Deployment Optimization of a Fleet of Drones for Routine Inspection of Networks of Linear Infrastructures

Abstract

In this paper, the problem of optimizing the deployment of a fleet of Unmanned Aerial Vehicles (UAVs) for the inspection of networks of linear infrastructures is addressed. In this optimization problem, two levels need to be considered: the UAVs need to follow a trajectory allowing them to fly over linear infrastructures, with a limited endurance, while ground vehicles need to be deployed to act as platforms that can launch and retrieve the UAVs from fixed parking positions. Both parking positions of ground vehicles and the UAVs fleet routes along the linear infrastructures have to be optimized to improve the inspection efficiency, which is related to multiple contradictory objectives such as minimizing operational cost and maximizing inspection performance. The algorithm proposed in this paper was developed by expanding upon several algorithms from literature. More precisely, we propose to solve a Two-Echelon Vehicle Routing Problem (2E-VRP) with a custom multi-objective Hybrid Genetic Algorithm (HGA) based on a Capacitated Arc Routing Problem (CARP). Simulation results are provided on a benchmark of networks of linear infrastructures along with a sensitivity analysis on the parameters of the algorithm, showing promising results.